Here are five worth revisiting, what each one actually targeted, and what to do about it now that the web is drowning in AI-generated everything.
Mobilegeddon
Back in April 2015, Google started using mobile-friendliness as a ranking factor for mobile searches. At the time, it felt like a nudge. It’s now the default lens Google uses to crawl and score almost everything.
The metric that measures this has changed shape since then, too. Interaction to Next Paint replaced First Input Delay as the responsiveness signal inside Core Web Vitals, and the current “good” thresholds sit around 200 milliseconds for INP, 2.5 seconds for Largest Contentful Paint, and 0.1 for Cumulative Layout Shift. None of that is a nice-to-have anymore, and it’s worth remembering that AI Overviews build their summaries off how a page actually renders, mobile first.
I’ve had clients pass every “mobile-friendly” checker with a clean bill of health and still get outperformed by a competitor, because the checker doesn’t catch a 4MB hero image choking a mid-range Android phone on real-world connection speeds.
Do this: Pull your field data from the Core Web Vitals report in Search Console, filtered to mobile, not lab data from your own fast office wifi. Fix the worst-performing fifth of your pages before you touch anything else this quarter.
Quality Updates
This traces back to the Panda-era pattern that started in 2011: demote sites stuffed with thin, duplicate, or scraped content. It’s probably the most important entry on this list, because it never really ended. It moved house.
Google folded its Helpful Content signal directly into the core ranking algorithm back in March 2024, so there’s no separate “helpful content update” left to wait for or recover from. Every core update since, including the ones that shipped this March and May, is evaluating the same underlying question: does this page sound like someone who actually knows the subject, or does it read like something assembled to rank. Google itself has said this shift meaningfully cut the amount of low-quality, unoriginal content showing up in results.
Do this: Run an honest content inventory. If a page can’t point to a specific data point, a named source, a photo you took yourself, or a stated opinion nobody else in your space is willing to say out loud, rewrite it with that detail or pull it down. This gets evaluated site-wide, so a pile of weak pages can drag down the strong ones sitting next to them.
Fred
Fred was the unofficial name (later confirmed by Google’s own team) for a 2017 update targeting sites loaded with aggressive ads and thin affiliate content built to serve ad impressions rather than readers.
The AI content mill is basically Fred’s original target audience with a faster printing press. Google’s spam policies now explicitly define “scaled content abuse,” publishing large volumes of pages mainly to manipulate rankings, regardless of whether a person or a model wrote them, as active grounds for removal from search results. If your growth plan is forty posts a week with three ad units stacked above the fold, that’s precisely the pattern Fred was built to catch, just with sharper detection pointed at you now.
Do this: Work out your ad-to-content ratio and your publishing pace, honestly. If you can’t defend a page’s depth against how much ad real estate sits on it, cut the volume before the algorithm does it for you.
EMD (Exact Match Domain)
Back in 2012, Google stopped handing out a free ranking boost just because a domain name matched a search term exactly, when the content sitting behind it was weak.
Easy to write off as ancient history. But the instinct behind it lives on in Google’s expired domain abuse policy, which targets people buying up old domains with real backlink history and repopulating them with unrelated, often AI-generated content to borrow authority they never earned. Same shortcut, newer packaging.
Do this: If you’re sitting on a domain you bought or inherited for its keyword or its history, check what’s actually published there now against what earned that authority in the first place. A great domain name or backlink profile buys you nothing if today’s content doesn’t live up to it.
Venice Update
This one, from 2012, made Google noticeably better at reading implicit local intent, showing local results for a query even when nobody typed “near me.”
Local and general results have kept blending ever since, and AI Overviews have sped that up, pulling in local business signals for searches that never mention a location at all. If you run a service business and treat your Google Business Profile as an afterthought because you’re “doing real SEO” elsewhere, you’re ignoring a signal that increasingly decides whether you show up at all.
Do this: Check whether your service and category pages carry actual location detail, not just an address buried in the footer, and keep your Business Profile categories, hours, and services genuinely current. A stale profile costs you more visibility now than it did a few years back.
Where this leaves you
Google doesn’t bring old updates out of retirement. It just keeps their logic running quietly and stacks new signals on top every few months. If your traffic has been sliding through 2025 and into this year, the useful exercise isn’t hunting for some brand-new 2026 update to pin the blame on. It’s going back through this list, page by page, and checking whether you ever actually fixed what these five already flagged.